Finding the Normal on a Cylinder
Once you know the points of intersection, the normal vector is used to help shade the surface appropriately. You’ll only need one scenario to cover this bit.
Test #3: Normal Vector on a Cylinder
Show that the normal vector on the surface of a cylinder is computed correctly.
This scenario chooses four points on the surface of the cylinder, one each at +x, -x, +z and -z, and shows that the normal is the expected value at each point.
| | Scenario Outline: Normal vector on a cylinder |
| | Given cyl ← cylinder() |
| | When n ← local_normal_at(cyl, <point>) |
| | Then n = <normal> |
| | |
| | Examples: |
| | | point | normal | |
| | | point(1, 0, 0) | vector(1, 0, 0) | |
| | | point(0, 5, -1) | vector(0, 0, -1) | |
| | | point(0, -2, 1) | vector(0, 0, 1) | |
| | | point(-1, 1, 0) | vector(-1, 0, 0) | |
To accomplish this, take the point in question and remove the y component. Treating the result as a vector gives you the normal. In pseudocode, it looks like this:
| | function local_normal_at(cylinder, point) |
| | return vector(point.x, 0, point.z) |
| | end function |
With those tests passing, your ray tracer can render cylinders! They’ll be infinitely long, which might be a bit unwieldy, but with a bit of imagination you can do all kinds of interesting things with them. Give it a try! When you come back, we will look at truncating those cylinders to make them easier to use.